RGI Desert Ghost Elimination Test

Ok. I’m confused by the testing. Here is what is stated.

What is the Desert Ghost Elimination Test?

The RGI test for Desert Ghost in Ball Pythons is an elimination test. It is 97% accurate for determining that an animal does NOT have the mutation, but it is only 85% accurate in determining that it does have the mutation. Further research is being done to improve this test to 99% accuracy in all scenarios.

So…if I send in a shed for test it come back negative. Is that negative for the gene or negative for eliminating that it has the gene. Maybe I’m confusing myself on this one.

Breeder I bought the girls from told me the Sire is one he bought as DG so the girls should be DG even though he wasn’t sure as he never paired the Sire with a het/homogenous to prove out.


Desert Ghost has turned out to be polygenic. So far they discovered what they’re calling A and B on two different chromosomes. And a copy of each is needed to create a visual. There’s at least one more gene they’re still trying to locate.

The test can tell you for sure if there are no A or B DG genes. It can tell if the animal is Het or Hom for A or B. But if it is not A or B, there’s a chance of false negative from the unknown gene. Once the others gene is located the test will become much simpler.

There’s a great write up here:


I understand that DG is polygenic. I’ve seen the vids they pushed out on the subject. I’m more confused on the wording of the test itself.


I believe if a DG teat comes back Negative then they are not DGa or DGb (to a 97% accuracy). There is a 3% chance the test is wrong.

If it comes back positive, there is only an 85% chance the test:

  • correctly IDed the strain of DG
  • and was within the 97% accuracy of the test itself.

Thanks. It’s just confusing how they can be positive on not having it and somewhat positive if it does.


That because they are looking for something specific, but not specific enough.

Lets say a detective in the 1920s is looking over a crime scene.
You and your 9 friends are suspects. 2 of your friends are ginger.

If the detective finds no ginger hair at the scene, then they can be pretty sure your ginger friends are not the culprits… 97% sure. A 3% room for human error.

But if they find a ginger hair, they are 85% sure that one of your ginger friends is the culprit, but they dont know which one. They’re accurate about identifying the connection and staying within the 97% accuracy of their detective skills, but only to a certain extent—85% of the time. So, it’s like the detective sometimes gets a bit excited and points to the right suspect, but there’s still a chance of a mistaken identity due to how similar your friends hair is to each other.


Yea…that doesnt help much. I understand they are looking for certain codes in the DNA. If those codes do not show up then it should be negative. If the codes do show up it should be positive. So realisticly their testing is only 85% accurate, which is still high but still has 15% error rate.


Please let me confirm all this with Shawn to be sure before being quoted…

Indeed, and in the majority of tests that is the case. But DG is a little different until a test for each “line” has been concreted. A DGd, e, f… could still be in the mix and causing inconsistencies.

This is basically RGI covering their own back if it turns out years later that there was another ginger haired person in the area that was not immediately identify as a suspect.

Ginger haired people, I am sorry to use you as an example :stuck_out_tongue_closed_eyes::heart:


I’m a ginger beard… yeah. DG is confusing for sure


Also, I’m pretty sure that was the original description of the test and may well be due for an update. At least the way I remember it (not sure when RGI knew what) that description came out before their podcast explaining DGa and DGb and the ongoing search for DGc etc. 85% might have referred to how just looking at one of the two (DGa or DGb) or even once they had tests for both how that still didn’t fully explain all the rules around when DG is visual.